Deep Research orchestrates a multi-source investigation across web search, SEC filings, earnings transcripts, research articles, and financial data. It scopes the question, collects evidence in parallel, compares sources to flag single-source claims, and produces a structured analyst-style brief.
更新于—
内置技能
Driven
内置技能
Before any data is pulled, the core question, time horizon, and market coverage are fixed. Scoping converts a vague prompt like "research NVDA" into a targeted plan — company vs industry lens, recent-quarter vs long-term trend, single market vs cross-market — so downstream collection stays focused instead of exhaustive.
Evidence is gathered in parallel across web search (neural for broad queries, keyword for precise ticker and phrase matches), article and transcript retrieval, SEC filings, earnings call transcripts, and structured financial data. No single source is trusted alone; the breadth is the point.
Claims are compared across sources — anything that lands from only one source is flagged as unverified. Facts (earnings numbers, price data) are separated from opinions (ratings, commentary); stale items are flagged by publication date.
Findings are collapsed into a structured report covering key findings, recent developments, financial snapshot, market sentiment, risks, and cited sources — delivered as a downloadable file so the chat stays clean and the full analysis is reviewable.